This paper demonstrates that multiple comparison tests using a pooled error term are dependent on the circularity assumption and shows how to compute tests which are insensitive (robust) to this assumption. (Author/GK)

If the null hypothesis of a one-sample test of multivariate means is rejected, the dimension of the line joining the population centroid and the hypothesized centroid can be interpreted with a linear function, using a discriminant function and the correlation of each dependent variable with a discriminant score. (Author/BW)

The validity conditions for univariate repeated measures designs are described. Attention is focused on the sphericity (equality of variance) requirement. It is recommended that separate rather than pooled error term procedures be routinely used to test a priori hypotheses. (Author/JKS)

Procedures for the analysis of profiles of means in repeated measures designs under order restriction for patterns of mean change are described. Tables of critical values are provided for the case of simple-order alternatives. (Author/JKS)

Cattell's scree test and Bartlett's chi-square test for the number of factors to be retained from a factor analysis are shown to be based on the same rationale, with the former reflecting subject sampling variability, and the latter reflecting variable sampling variability. (Author/JKS)

The popular factor analytic procedure of putting unities in the main diagonal and retaining all factors with an eigenvalue greater than one is criticized with respect to the number of factors retained and resultant estimates of common variance. Ways of avoiding these difficulties are discussed. (JKS)

Monte Carlo techniques were used to compare the power of Wilcoxon's rank-sum test to the power of the two independent means t test for situations in which samples were drawn from (1) uniform, (2) Laplace, (3) half-normal, (4) exponential, (5) mixed-normal, and (6) mixed-uniform distributions. (Author/JKS)

Errors of misclassification associated with two concept acquisition criteria and their effects on the actual significance level and power of a statistical test for sequential development of these concepts are presented. Explicit illustrations of actual significance levels and power values are provided for different misclassification models.…

The effect of the variance stabilizing transformations on the significance level and power of the F test, applied to Binomial and Poisson variables, was examined by a Monte Carlo study. The results indicate that the significance level is not affected by the transformations. (Author/JKS)

Procedures are developed for testing the hypothesis that Cronbach's alpha reliability coefficient is equal for two tests given to the same subjects. (Author/JKS)

Orthogonal target analysis, a technique employed in confirmatory factor analysis, is investigated via a simulation study. The results indicate that the technique will recover the correct underlying population pattern except under very unfavorable data conditions and that a close fit to a binary target is not necessarily forced. (Author/JKS)

Detailed reanalyses of data reported in Studies in Educational Evaluation: Monograph No. 1 by Y. Kashti and Monograph No. 5 by U. Kattmann, 1979, were performed using an explicitly structurally oriented approach via target analysis (PINDIS). Results contradict those reached by Kashti and Kattmann. (RL)

The authors contend that most research in education lacks statistical power. They feel that the poor use of statistics as a tool for enhancing internal validity must be remediated. The adoption of a new convention is proposed in order to put statistical certainty into reasonable perspective. (RLV)

It has been shown that the degree of restriction of range taken into account in testing the hypothesis that rho equals zero, entails risks of incorrect inferences. It is argued that an alternative is to disregard the restriction of range and to use the common t-statistics proposed by regression theory. (Author/JKS)

A program is described which selects the most powerful among four methods for conducting a priori comparisons in an analysis of variance: orthogonal contrasts, Scheffe's method, Dunn's method, and Dunnett's test. The program supplies the critical t ratio and the per-comparison Type I error risk for each of the relevant methods. (Author/JKS)